JP7404694B2 - Relevance evaluation method, relevance evaluation device, program - Google Patents

Description

The present invention relates to a relevance evaluation method, a relevance evaluation device, and a program.

Calculate and evaluate the relationship between people and words in order to efficiently search for people related to a specific technology, field, product, project, etc. in organizations such as companies and organizations. There are things to do.

As a technique used when calculating such a relationship, there is, for example, Patent Document 1. Patent Document 1 describes a related score calculation system that includes a collection unit, a word extraction unit, and a related score calculation unit. According to Patent Document 1, a word extraction unit extracts words from file names written in operation logs collected from terminal devices by a collection unit. Further, the association score calculation unit calculates an association score representing the strength of association between the user and the word based on the operation log. Furthermore, Patent Document 1 describes two methods for calculating relevance scores: a first method in which the relevance is quantified for each word for one user, and another method in which the relevance is quantified for each word for one user. Two methods are described: a second method that digitizes the relationship, which also takes into account the degree of association between words.

JP2019-86940A

When relevance is quantified using the second method described in Patent Document 1, for example, it is determined that a person who uses the word "smartphone" is also related to the word "smartphone". It is expected that it will be more resistant to spelling fluctuations. However, in the case of the second method, for example, if user A often uses the word "B" and there is a tendency among all users who use the word "A" to also often use the word "B". , even if user A did not use the word "A", the association score for user A with "A" might also be high.

As described above, when the second method, which is more resistant to synonyms and spelling variations, is used, there is a problem in that the association score for words that are not actually used may become high. These problems were more pronounced with highly specialized words that were used by fewer users.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a relevance evaluation method that solves the problem that it is difficult to suppress the influence of high relevance scores for words that are not actually used, while being robust against synonyms and spelling variations. Our goal is to provide sex evaluation devices and programs.

In order to achieve this purpose, a relevance evaluation method that is one form of the present invention includes:
The information processing device
calculating a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
Calculating a second score based on the first score and the degree of association between words,
The configuration is such that information for identifying users is rearranged based on the calculated first score and second score.

Further, a relevance evaluation device according to another embodiment of the present invention includes:
a first score calculation unit that calculates a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score based on the first score and the degree of association between words;
a sorting processing unit that sorts information for identifying users based on the calculated first score and second score;
It has the following structure.

Further, a program which is another form of the present invention is
In the information processing device,
a first score calculation unit that calculates a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score based on the first score and the degree of association between words;
a sorting processing unit that sorts information for identifying users based on the calculated first score and second score;
This is a program to achieve this.

By having the above configuration, the present invention solves the problem that it is difficult to suppress the influence of high association scores for words that are not actually used, while making it resistant to synonyms and spelling variations. It becomes possible to provide a relevance evaluation method, a relevance evaluation device, and a program.

FIG. 1 is a block diagram showing an example of the configuration of a relevance evaluation device according to a first embodiment of the present invention. FIG. 2 is a diagram showing an example of reference time information shown in FIG. 1; 2 is a diagram showing an example of first score information shown in FIG. 1. FIG. 2 is a diagram showing an example of a first table shown in FIG. 1. FIG. FIG. 2 is a diagram showing an example of a second table shown in FIG. 1; It is a figure which shows an example of the 2nd score information shown in FIG. 2 is a diagram showing an example of a ranking table included in the ranking information shown in FIG. 1. FIG. 2 is a diagram showing another example of a ranking table included in the ranking information shown in FIG. 1. FIG. FIG. 3 is a diagram for explaining an example of a process performed by a first score calculation unit. FIG. 7 is a diagram for explaining an example of processing performed by a second table generation unit. It is a flowchart which shows an example of operation of a relevance evaluation device. 3 is a flowchart illustrating an example of processing for generating a ranking table. It is a flow chart which shows an example of operation when a relevance evaluation device performs a search. It is a block diagram showing an example of composition of a relevance evaluation device in a 2nd embodiment of the present invention.

[First embodiment]
A first embodiment of the present invention will be described with reference to FIGS. 1 to 13. FIG. 1 is a block diagram showing an example of the configuration of the relevance evaluation device 100. FIG. 2 is a diagram showing an example of the reference time information 111. FIG. 3 is a diagram showing an example of the first score information 112. FIG. 4 is a diagram showing an example of the first table 113. FIG. 5 is a diagram showing an example of the second table 114. FIG. 6 is a diagram showing an example of the second score information 115. 7 and 8 are diagrams showing examples of ranking tables included in the ranking information 116. FIG. 9 is a diagram for explaining an example of a process performed by the first score calculation unit 120. FIG. 10 is a diagram for explaining an example of processing performed by the second table generation unit 140. 11 to 13 are flowcharts showing an example of the operation of the relevance evaluation device 100.

In the first embodiment of the present invention, the strength of the relationship between a person and a word, such as a certain person spending a lot of time using a certain word, is quantified and displayed in a ranking format. The relevance evaluation device 100 that evaluates the strength of gender will be described. As will be described later, the relevance evaluation device 100 determines whether a word is related to another word that has a similar distribution of users and a first score that quantifies the relevance of each word to each user. A second score is calculated by adding the degree of association between the first score and the first score. Then, the relevance evaluation device 100 generates a ranking based on the calculated first score, second score, and word usage status such as the number of users who have used the word.

The relevance evaluation device 100 is an information processing device that evaluates the strength of the relevance between a person and a word by calculating the above-described first score and second score and generating a ranking. Furthermore, the relevance evaluation device 100 is configured to be able to perform a search using the generated rankings.

FIG. 1 shows an example of the configuration of a relevance evaluation device 100. Referring to FIG. 1, the relevance evaluation device 100 includes, for example, a storage unit 110, a first score calculation unit 120, a first table generation unit 130, a second table generation unit 140, and a second score calculation unit 150. , a ranking generation section 160 , a keyword reception section 170 , a search section 180 , and an output section 190 .

For example, the relevance evaluation device 100 includes a calculation device such as a CPU (Central Processing Unit). For example, the relevance evaluation device 100 realizes each of the processing units described above by having a calculation device execute a program stored in a storage device such as the storage unit 110.

The storage unit 110 is a storage device such as a hard disk or a memory. The main information stored in the storage unit 110 includes, for example, reference time information 111, first score information 112, first table 113, second table 114, second score information 115, and ranking information 116. There is. Further, the storage unit 110 can store a program read in advance from an external device, a storage medium, or the like.

The reference time information 111 is information indicating the reference time, which is the time during which the user referred to the word, for each pair of user and word. The reference time information 111 is generated in advance and stored in the storage unit 110 based on, for example, file operation logs and schedule information. Examples of word reference times include the operation time of a user who opened a file that includes a certain word in the file name, the number of keyboard types, and the time from start to end of schedule information that includes a certain word in the subject, etc. . The reference time of the word may be other than those exemplified above.

FIG. 2 shows an example of the reference time information 111. As shown in FIG. 2, in the reference time information 111, for example, a user name, a word, and a reference time are associated with each other. For example, the first line in FIG. 2 shows that the reference time of the word "work" for the user name "Mr. A" is "30".

The first score information 112 is information indicating a first score that is a numerical representation of the relevance of each word for each user. That is, the first score information 112 includes a first score indicating the degree of association between the user and the word. The first score indicated by the first score information 112 is calculated based on the reference time information 111 by a first score calculation unit 120 described later.

FIG. 3 shows an example of the first score information 112. As shown in FIG. 3, in the first score information 112, for example, a user name, a word, and a first score are associated with each other. For example, the first line of FIG. 3 shows that the first score of the word "service" for the user name "Mr. A" is "0".

Note that the first score indicated by the first score information 112 has a characteristic that when the reference time of a word by the user is 0, the corresponding score value is 0, and the longer the reference time, the higher the score value. are doing. Further, the first score has a characteristic that the score value becomes lower for a word that is used by all users, and the score value becomes higher for a word that is used only by a specific user. Because of the above characteristics, the first score has a high value when the user and the word are evaluated to be highly related, such as when the user uses the word often or when only a specific user uses the word. It can be said that the score is

The first table 113 is a table generated by assigning each user to a row and each word to a column based on the first score information 112. The first table 113 is generated by a first table generation unit 130, which will be described later.

FIG. 4 shows an example of the first table 113. For example, in the first line of Figure 4, the first score of the word "service" of the user name "Mr. A" is "0", the first score of the word "vacation" is "0", and the first score of the word "vacation table" is "0". The first score for each word of the user name "Mr. A" is shown as "0", . . . .

The second table 114 is a table showing the degree of association between words. The second table 114 is generated based on the first score information 112 and the first score indicated by the first table 113 by a second table generation unit 140, which will be described later.

FIG. 5 shows the second table 114. For example, in the first line of FIG. ” and the word “work style” are “-0.65578”, etc., indicating the degree of association between the word “meeting” and other words.

Note that the degree of association between words indicated by the second table 114 is an index that increases in value for words that are used in similar ways, such as having similar first score distributions. For example, in the case of the degree of association between word x and word y, there are many people who use both word x and word y, or many people who do not use both word x and word y. The more similar they are, the larger the value. For example, it is assumed that the words ``smartphone'' and ``smartphone'' are used in similar ways, resulting in a large value.

The second score information 115 is information indicating a second score obtained by adding the degree of association between words to the first score, assuming that the word is also related to another word that has a similar distribution of users. The second score indicated by the second score information 115 is determined by a second score calculation unit 150 (described later) based on the degree of association between the first score indicated by the first score information 112 or the first table 113 and the word indicated by the second table 114. Calculated based on.

FIG. 6 shows an example of the second score information 115. As shown in FIG. 6, in the second score information 115, user names, words, and second scores are associated with each other. For example, the first line of FIG. 6 shows that the first score of the word "service" for the user name "Mr. A" is "0.145650".

Note that the second score indicated by the second score information 115 is calculated based on the first score and the degree of association between words, as described above. Therefore, by using the second score, for example, when performing a search described below, it is possible to search for user names of users that are highly related to another word that is highly related to the word specified as a search keyword. It becomes possible. In other words, it can be said that the second score indicated by the second score information 115 is a value that is more resistant to synonyms and spelling variations than the first score.

Ranking information 116 indicates the degree of relevance of a person to a word in a ranking format for each word. As shown in FIGS. 7 and 8, the ranking information 116 includes a ranking table generated in the first format or the second format for each word. The ranking table included in the ranking information 116 is generated by a ranking generation unit 160, which will be described later. Furthermore, which of the first format shown in FIG. 7 and the second format shown in FIG. 8 should be used to generate the ranking table is determined for each word by the ranking generation unit 160, which will be described later.

The ranking table in the first format shown in FIG. 7 is generated as a result of sorting the information identifying the users in descending order of the second scores, as will be described later. As shown in FIG. 7, in the case of the ranking table in the first format, the higher the value of the second score, the higher the ranking, and the first score is not taken into consideration. For example, in the second line of FIG. 7, "Mr. B", whose second score is "0.043959" and first score is "0", is associated with the user ranked "2".

On the other hand, in the ranking table in the second format shown in FIG. 8, the information that identifies users is sorted in descending order of the first score, and then the users whose first score is 0 are sorted in the descending order of the second score. generated as a result of As shown in Figure 8, in the case of the ranking table in the second format, the higher the value of the first score, the higher the ranking, and the value of the second score has an effect only when the value of the first score is 0. There is. For example, in the second line of FIG. 8, "Mr. C", who has a second score of "0.024045" and a first score of "0.003854", is associated as a user ranked "2".

Note that the information for specifying a user included in the ranking table includes, for example, a user name. The information specifying the user may include the first score and the second score, or may include various other information.

The first score calculation unit 120 calculates, for each user, a first score for each word based on the reference time information 111, which is a numerical representation of the relationship between the user and the word. The first score calculation unit 120 then stores the calculated first score in the storage unit 110 as the first score information 112.

For example, the first score calculation unit 120 calculates the first score indicating the relationship between the user and the word by calculating the TF-IDF shown in FIG. Here, TF-IDF is a method for evaluating the importance of words included in a document. The first score calculation unit 120 calculates TF-IDF as a first score by considering the words used by each user as sentences and applying them to the calculation formula shown in FIG. The calculation procedure by the first score calculation unit 120 is as follows.

First, the first score calculation unit 120 refers to the reference time information 111 and totals the reference times of all words for each user, thereby obtaining a first total that indicates the total reference time of all words for a certain user. Calculate reference time. In addition, the first score calculation unit 120 calculates a second total reference time indicating the total reference time of all users for a certain word by summing the reference time of all users for each word.

Further, the first score calculation unit 120 calculates TF by dividing the reference time of a certain word by a certain user by the first total reference time, and also divides the reference time of a certain word by a certain user by the second total reference time. Calculate IDF by Then, the first score calculation unit 120 calculates TF-IDF for a certain user and a certain word by multiplying the calculated TF and IDF.

For example, the first score calculation unit 120 calculates the TF-IDF (that is, the first score) of each user and each word by performing the above-described process for each combination of user and word. The first score calculation unit 120 then stores the calculated result in the storage unit 110 as the first score information 112. Note that if the reference time of a certain word by a certain user is 0, the values of TF and IDF are both 0. Therefore, the value of the first score for a certain user and a certain word will also be 0.

The first table generation unit 130 generates the first table 113 based on the first score information 112. For example, the first table generation unit 130 generates the first table 113 by assigning each user included in the first score information 112 to a row and each word to a column. The first table generation unit 130 then stores the generated first table 113 in the storage unit 110.

The second table generation unit 140 generates a second table 114 indicating the degree of association between words based on the first score indicated by the first score information 112 or the first table 113. Then, the second table generation unit 140 stores the generated second table 114 in the storage unit 110.

For example, the second table generation unit 140 calculates the degree of association between words by using the formula for calculating the degree of association r shown in FIG. Here, the formula shown in Figure 9 is the covariance (Sxy) of the first scores of both words for n users, assuming that the total number of users is n and two words x and y are selected from the target words. The degree of association between words is calculated from the standard deviation (Sx, Sy) of the first scores of n people for each word.

Note that xi is the i-th first score for word x, and yi is the i-th first score for word y. Moreover, the following number 1 in the calculation formula shown in FIG. 10 is the average value of the first score of x, and the following number 2 in the calculation formula shown in FIG. 10 is the average value of the first score of y.

Note that the degree of association r shown in FIG. 10 is an index that increases in value for words that are used in similar ways. In other words, the degree of association between words increases when there are many people who use both word x and word y, or when there are many people who do not use both word x and word y. For example, "smartphones" and "smartphones" are used in similar ways, and their values are likely to be large.

The second table generation unit 140 calculates the degree of association between each word using the above-described calculation formula. Then, the second table generation unit 140 creates a second table 114 based on combinations of words so that the degree of association between each word is in the form of a target matrix. Note that the degree of association between words with itself is set to 1.

The second score calculation unit 150 calculates the degree of association between words based on the first score indicated by the first score information 112 or the first table 113 and the degree of association between words indicated by the second table 114. A second score is calculated by taking the score into consideration. Then, the second score calculation unit 150 stores the calculated second score in the storage unit 110 as second score information 115.

For example, the second score calculation unit 150 calculates the second score using the following formula.
Association score between user α and word X = Σ (degree of association between word X and a certain word x user A's first score for a certain word)

Specifically, for example, when calculating the second score of the user name "Mr. A" and the word "service" using the first table 113 shown in FIG. 4 and the second table 114 shown in FIG. The second score calculation unit 150 calculates the second score by performing the following calculation. In other words, the second score calculation unit 150 calculates (degree of association between "service" and "meeting" x first score of "Mr. A's" "meeting") + (degree of association between "service" and "work" x "Mr. A's" first score of "work") + (degree of relationship between "service" and "work style" x "Mr. A's" first score of "work style") + ("service" and "restraint" ” × ``Mr. A'''s first score of ``restraint'') + (degree of relationship between ``service'' and ``reform'' × ``Mr. A'''s 1st score of ``reform'') + (``Mr. A'''s 1st score of ``reform'') ” and “overtime work” x “Mr. A”’s first score of “overtime”) + (degree of association between “service” and “overtime work control” x “Mr. A”’s first score of “overtime work control”) score) + (degree of association between “service” and “service” × first score of “service” for “Mr. A”) + (degree of association between “service” and “vacation” × “vacation” of “person A”) ' first score) + (degree of association between 'service' and 'vacation table' x first score of 'A's' 'vacation table') = (-0.46217 x 0) + (0.941227 x 0.063492) + ( 0.941227 x 0.063492) + (-0.40357 x 0.027778) + (0.941227 x 0.063492) + (-0.40357 x 0.027778) + (-0.40357 x 0.027778) + (1 x 0) + (1 x 0) + (1 x 0) = Calculate 0.145650.

The ranking generation unit 160 generates a ranking table for each word, based on the first score information 112 and the second score information 115, indicating the degree of human relevance to the word. The ranking table generated by the ranking generation unit 160 has either a first format as shown in FIG. 7 or a second format as shown in FIG. 8. The ranking generation unit 160 then stores the generated ranking table in the storage unit 110 as the ranking information 116.

For example, the ranking generation unit 160 measures the number of users whose first score is not 0 among the users who have calculated the first score for the target word. In other words, the ranking generation unit 160 measures the number of users who have used the word.

If the measured number of users is equal to or greater than a predetermined threshold, the ranking generation unit 160 generates a ranking table in the first format by rearranging the information identifying the users in descending order of the second scores. On the other hand, if the number of measured users is less than the threshold, the ranking generation unit 160 rearranges the information identifying the users in descending order of the first score, and then assigns the users whose first score is 0 to the second score. By rearranging in descending order, a ranking table in the second format is generated. In this way, the ranking generation unit 160 rearranges information specifying users in different formats depending on the number of users measured. Further, in the first format, sorting is performed based on the second score, and in the second format, sorting is performed based on the first score and the second score.

Note that the threshold value described above may be any value. Further, the threshold value may be other than a predetermined fixed value. For example, the threshold value may be expressed as a percentage such as n% of the total number of people. Further, the threshold value may be a value that is dynamically changed based on feedback from the user.

In this way, the ranking generation unit 160 determines which ranking table to generate, the first format or the second format, based on the usage status of the word, such as the number of users who have used the word. . The ranking generation unit 160 then generates a ranking table in the determined format.

Note that the ranking generation unit 160 generates the ranking table as described above for each word included in the reference time information 111 and the like. Therefore, the ranking information 116 includes a ranking table for each word in either the first format or the second format.

The keyword accepting unit 170 accepts search keywords from searchers. For example, the keyword accepting unit 170 accepts words as search keywords.

The search unit 180 executes a search according to a search keyword. For example, the search unit 180 searches the ranking table of the word indicated by the search keyword from the ranking information 116.

The output unit 190 outputs a ranking table that is the search result by the search unit 180. The output by the output unit 190 includes, for example, display on a screen display unit and transmission to an external device.

As an example of implementing a search by the keyword reception unit 170, search unit 180, and output unit 190, for example, the keyword reception unit 4 receives a search keyword from an external device used by a searcher via a communication network. The output unit 190 then outputs the search results to an external device used by the searcher. Such a format is possible. The search may be performed by accepting a search keyword using a keyboard or the like included in the relevance evaluation device 100 and displaying the search keyword on a screen display unit included in the relevance evaluation device 100. Further, the keyword accepting unit 170 may be configured to extract words from a given sentence and accept the extracted words as search keywords, for example. In other words, the keyword accepting unit 170 does not necessarily have to be configured to directly accept input of a search keyword from a searcher. Note that in this embodiment, the content of the process performed when the keyword reception unit 170 extracts words from a sentence is not particularly limited. The keyword reception unit 170 can be configured to extract words from sentences using known techniques.

The above is an example of the configuration of the relevance evaluation device 100. Next, an example of the operation of the relevance evaluation device 100 will be described with reference to FIGS. 11 to 13.

First, with reference to FIG. 11, an example of the operation of the relevance evaluation device 100 when generating a ranking table and storing it as the ranking information 116 will be described.

Referring to FIG. 11, the first score calculation unit 120 of the relevance evaluation device 100 calculates, for each user, a first score calculation unit 120 that quantifies the relevance between the user and the word for each word based on the reference time information 111. A score is calculated (step S101). For example, the first score calculation unit 120 calculates the first score by calculating the TF-IDF shown in FIG. 8.

The first table generation unit 130 generates the first table 113 based on the first score information 112 (step S102).

The second table generation unit 140 generates a second table 114 indicating the degree of association between words based on the first table 113 (step S103). For example, the second table generation unit 140 calculates the degree of association between words by calculating the calculation formula shown in FIG. The second table generation unit 140 then generates the second table 114 by expressing the degree of association between the generated words in a table format.

The second score calculation unit 150 calculates the degree of association between words based on the first score indicated by the first score information 112 or the first table 113 and the degree of association between words indicated by the second table 114. A second score is calculated by taking the score into consideration (step S104). For example, the second score calculation unit 150 calculates the degree of association between word X and a certain word x the first score of user A for a certain word for each word, and calculates the sum, thereby obtaining the second score. Calculate.

The ranking generation unit 160 generates, for each word, a ranking table indicating the degree of human relevance to the word, based on the first score information 112 and the second score information 115 (step S105). Whether the ranking generation unit 160 generates the ranking table in the first format or the second format is determined based on, for example, the usage status of the word.

The above is an example of the operation of the relevance evaluation device 100 when generating a ranking table and storing it as the ranking information 116. Next, with reference to FIG. 12, the ranking generation process in step S105 will be described in more detail.

Referring to FIG. 12, the ranking generation unit 160 measures the number of users whose first score is not 0 for the target word (step S201).

If the measured number of users is equal to or greater than a predetermined threshold (step S202, Yes), the ranking generation unit 160 sorts the information identifying the users in descending order of the second score (step S203). Thereby, the ranking generation unit 160 generates a ranking table in the first format.

On the other hand, if the measured number of users is less than the predetermined threshold (step S202, No), the ranking generation unit 160 sorts the information identifying the users in descending order of the first score (step S204). Furthermore, the ranking generation unit 160 sorts the users whose first score is 0 in descending order of their second scores (step S205). Thereby, the ranking generation unit 160 generates rankings in the second format.

The above is an example of the ranking generation process shown in step S105 in FIG. 11. Next, with reference to FIG. 13, an example of the operation of the relevance evaluation device 100 when performing a search process will be described.

Referring to FIG. 13, the keyword accepting unit 170 accepts a search keyword from a searcher (step S301). For example, the keyword accepting unit 170 accepts words as search keywords.

The search unit 180 executes a search according to the search keyword (step S302). For example, the search unit 180 searches the ranking table of the word indicated by the search keyword from the ranking information 116.

The output unit 190 outputs a ranking table that is the search result by the search unit 180 (step S303). The output by the output unit 190 includes, for example, display on a screen display unit and transmission to an external device.

The above is an example of the operation of the relevance evaluation device 100 when performing a search process.

In this way, the relevance evaluation device 100 includes the first score calculation unit 120 that calculates the first score, the second score calculation unit 150 that calculates the second score, and the ranking generation unit 160. . With such a configuration, the ranking generation unit 160 can perform sorting based on the second score according to word usage, etc., and a second format sorting based on the first and second scores. It is possible to decide which format to generate the ranking table in. This makes it possible to deal with variations in spelling by sorting using the second score, and to rank actual users higher when there are fewer users of a word, such as a highly specialized word. In other words, according to the above configuration, it is possible to suppress the influence of high association scores for words that are not actually used, while making it resistant to synonyms and spelling variations.

In addition, in this embodiment, the case where the relevance evaluation apparatus 100 is comprised by one information processing apparatus was illustrated. However, the relevance evaluation device 100 may be configured by, for example, a plurality of information processing devices connected via a network. For example, the relevance evaluation device 100 includes an information processing unit 110, a first score calculation unit 120, a first table generation unit 130, a second table generation unit 140, a second score calculation unit 150, and a ranking generation unit 160. The information processing apparatus may include an information processing apparatus having a keyword reception section 170, a search section 180, and an output section 190.

Furthermore, in this embodiment, a case has been exemplified in which a ranking table is generated and stored in the ranking information 116 in advance. However, the ranking table may be configured to be generated, for example, when the search unit 180 performs a search. In this way, the generation timing of the ranking table is not limited to the case illustrated in this embodiment.

Furthermore, in the present embodiment, a case has been exemplified in which it is determined in which format the ranking table is generated based on the result of counting the number of users whose first score is not 0. However, the ranking table generation unit 160 may be configured to determine which format to generate the ranking table in, the first format or the second format, by a method other than the one exemplified above. For example, the ranking generation unit 160 may be configured to determine in which format the ranking table is generated based on the number of users whose first score is equal to or higher than a predetermined reference threshold. In other words, the ranking generation unit 160 may be configured to generate a ranking table in the first format when the number of users whose first score is greater than or equal to a predetermined reference threshold is greater than or equal to a predetermined threshold. I can do it. Furthermore, the ranking generation unit 160 selects one of the formats based on, for example, the number of users whose first score is less than a predetermined reference threshold and who are present in a predetermined rank or higher when the ranking table in the first format is generated. It may be configured such that it is determined whether to generate a ranking table. Further, the ranking generation unit 160 may be configured to take into account the situation of the second score and the like to determine in which format the ranking table is generated. Note that the reference threshold value may be any value.

Furthermore, the process of calculating the degree of association between words by the second table generation unit 140 requires a large amount of calculation depending on the number of words. Therefore, the number of words for which the degree of association is calculated may be limited by a method such as setting a usage rate threshold value for the usage rate.

[Second embodiment]
Next, referring to FIG. 14, a second embodiment of the present invention will be described. In the second embodiment, an overview of the configuration of the relevance evaluation device 20 will be described.

FIG. 14 shows an example of the configuration of the relevance evaluation device 20. Referring to FIG. 13, the relevance evaluation device 20 includes a first score calculation section 21, a second score calculation section 22, and a rearrangement processing section 23.

For example, the relevance evaluation device 20 includes a calculation device such as a CPU and a storage device. For example, the relevance evaluation device 20 realizes each of the processing units described above by having a calculation device execute a program stored in a storage device.

The first score calculation unit 21 calculates a first score indicating the degree of association between the user and the word, based on information indicating the reference time of the word.

The second score calculation unit 22 calculates a second score based on the first score calculated by the first score calculation unit and the degree of association between words.

The rearrangement processing unit 23 rearranges the information identifying the users based on the first score calculated by the first score calculation unit 21 and the second score calculated by the second score calculation unit 22.

In this way, the relevance evaluation device 20 includes a first score calculation section 21 , a second score calculation section 22 , and a rearrangement processing section 23 . With such a configuration, the sorting processing unit 23 sorts the information identifying the user based on the first score calculated by the first score calculation unit 21 and the second score calculated by the second score calculation unit 22. You can make changes. As a result, the sorting processing unit 23 can determine, for example, whether to perform sorting based on the second score or based on the first score and second score. This makes it possible to suppress the effects of high association scores for words that are not actually used, while making it resistant to synonyms and spelling variations.

Furthermore, the above-described relevance evaluation device 20 can be realized by incorporating a predetermined program into the relevance evaluation device 20. Specifically, a program according to another embodiment of the present invention causes an information processing device to calculate a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word. A score calculation unit 21, a second score calculation unit that calculates a second score based on the first score and the degree of association between words, and a user identified based on the calculated first score and second score. This is a program for realizing a rearrangement processing unit that rearranges information to be processed.

Further, in the relevancy evaluation method executed by the relevancy evaluation device 20 described above, the information processing device calculates a first score indicating the degree of relevancy between the user and the word based on information indicating the reference time of the word. and calculating a second score based on the first score and the degree of association between words, and rearranging information that identifies the user based on the calculated first score and second score. This is the method.

Even if the invention is a program or a relevance evaluation method having the above-mentioned configuration, it is possible to achieve the above-mentioned object of the present invention because it has the same operation and effect as the relevance evaluation device 20. I can do it.

<Additional notes>
Part or all of the above embodiments may also be described as in the following additional notes. Hereinafter, an outline of the relevance evaluation device and the like in the present invention will be explained. However, the present invention is not limited to the following configuration.

(Additional note 1)
The information processing device
calculating a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
Calculating a second score based on the first score and the degree of association between words,
A relevance evaluation method in which information for identifying users is rearranged based on the calculated first score and second score.
(Additional note 2)
The relevance evaluation method described in Appendix 1,
Determine the format for sorting based on word usage,
A relevance evaluation method that performs sorting according to a determined format.
(Additional note 3)
The relevance evaluation method described in Appendix 2,
A relevance evaluation method, wherein usage of a word is determined based on the number of users whose first score is equal to or greater than a predetermined reference threshold among the users who have calculated the first score.
(Additional note 4)
The relevance evaluation method according to any one of Supplementary Notes 1 to 3,
A first format that performs sorting based on the second score, and a second format that performs sorting based on the first score and the second score. Gender evaluation method.
(Appendix 5)
The relevance evaluation method described in Appendix 4,
In the first format, information for identifying users is sorted in descending order of the second score.
(Appendix 6)
The relevance evaluation method described in Appendix 4 or Appendix 5,
In the second format, after sorting the information that identifies the users in descending order of the first score, the information that identifies the users whose first score is 0 is sorted in the descending order of the second score. Sort by Relevance evaluation method.
(Appendix 7)
The relevance evaluation method according to any one of Supplementary notes 1 to 6,
A relevance evaluation method that searches for results sorted based on search keywords and outputs the search results.
(Appendix 8)
The relevance evaluation method according to any one of Supplementary Notes 1 to 7,
A relevance evaluation method that calculates a degree of relevance between words based on the first score.
(Appendix 9)
a first score calculation unit that calculates a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score based on the first score and the degree of association between words;
a sorting processing unit that sorts information for identifying users based on the calculated first score and second score;
A relevance evaluation device having:
(Appendix 10)
In the information processing device,
a first score calculation unit that calculates a first score indicating the degree of association between the user and the word based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score based on the first score and the degree of association between words;
a sorting processing unit that sorts information for identifying users based on the calculated first score and second score;
A program to achieve this.

Note that the programs described in each of the above embodiments and supplementary notes are stored in a storage device or recorded in a computer-readable recording medium. For example, the recording medium is a portable medium such as a flexible disk, an optical disk, a magneto-optical disk, and a semiconductor memory.

Although the present invention has been described above with reference to the embodiments described above, the present invention is not limited to the embodiments described above. The configuration and details of the present invention can be modified in various ways that can be understood by those skilled in the art within the scope of the present invention.

100 Relevance evaluation device 110 Storage unit 111 Reference time information 112 First score information 113 First table 114 Second table 115 Second score information 116 Ranking information 120 First score calculation unit 130 First table generation unit 140 Second table generation Section 150 Second score calculation section 160 Ranking generation section 170 Keyword reception section 180 Search section 190 Output section 20 Relevance evaluation device 21 First score calculation section 22 Second score calculation section 23 Sorting processing section

Claims (4)

The information processing device
Based on the information indicating the reference time of the word, calculate a first score that is a value according to the reference time of the word by the user and indicates the height of the relationship between the user and the word,
Based on the first score and the degree of association between words, a second score is calculated by adding the degree of association between words to the first score,
By sorting the information that identifies users based on the calculated first score, second score, and word usage, it shows the strength of the relationship between people and words, and makes it easier to search. generate a ranking of
When generating the ranking, for the word for which the ranking is to be generated, the number of users whose first score is equal to or higher than a threshold value is measured among the users who have calculated the first score, and the number of users whose first score is equal to or higher than a threshold value is measured. In one case, the ranking is generated by rearranging information that identifies users in descending order of the second score, and if the number of measured users is less than a threshold, the users are identified in descending order of the first score. 2. A relevance evaluation method, wherein the ranking is generated by sorting the information on the first score, and then sorting the users whose first score is 0 in descending order of the second score . The relevance evaluation method according to claim 1,
A relevance evaluation method that searches the rankings that are the results of sorting based on the obtained search keywords and outputs the search results. a first score calculation unit that calculates a first score, which is a value according to the user's reference time of the word and indicates a high degree of association between the user and the word, based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score by adding the degree of association between words to the first score based on the first score and the degree of association between words;
By sorting the information that identifies users based on the calculated first score, second score, and word usage, it shows the strength of the relationship between people and words, and makes it easier to search. a sorting processing unit that generates a ranking,
has
When generating the ranking, the sorting processing unit measures the number of users whose first score is equal to or higher than a reference value for the word for which the ranking is to be generated, among the users who have calculated the first score. If the number of users is greater than or equal to the threshold, the ranking is generated by rearranging the information that identifies the users in descending order of the second score, and if the number of users measured is less than the threshold, the first A relevance evaluation device that generates the ranking by sorting information specifying users in descending order of scores, and then sorting users whose first score is 0 in descending order of their second scores . In the information processing device,
a first score calculation unit that calculates a first score, which is a value according to the user's reference time of the word and indicates a high degree of association between the user and the word, based on information indicating the reference time of the word;
a second score calculation unit that calculates a second score by adding the degree of association between words to the first score based on the first score and the degree of association between words;
By sorting the information that identifies users based on the calculated first score, second score, and word usage, it shows the strength of the relationship between people and words, and makes it easier to search. a sorting processing unit that generates a ranking,
Realize,
When generating the ranking, the sorting processing unit measures the number of users whose first score is equal to or higher than a reference value for the word for which the ranking is to be generated, among the users who have calculated the first score. If the number of users is greater than or equal to the threshold, the ranking is generated by rearranging the information that identifies the users in descending order of the second score, and if the number of users measured is less than the threshold, the first A program that generates the ranking by sorting information specifying users in descending order of scores, and then sorting users whose first score is 0 in descending order of their second scores .

Images